This project component will focus on establishing these behaviors essential for demonstrating preference, making requests, and choosing among preferred stimuli in individuals with profound retardation. These are individuals who have a history either of not controlling their environment in any way or not learning simple contingencies. Two previous lines of research are the basis of this research. In one, individuals with profound retardation and other severe medical and sensory impairments were exposed to switches that controlled activation of leisure devices such as radios and vibrators. Some of these individuals showed higher rates of switch depression when these response activated particular leisure devices were activated or when no leisure device was activated. This differential responding suggested that these individuals developed volitional control over an aspect of their environment and a rudimentary way of communicating preferences to others. Others did not show differentiated responding which suggests that the contingency was not learned. In the second line of research, individuals with severe or profound retardation, but with more physical capabilities were taught a simple vocational task. These individuals initially showed no differences in rates of work under a rich fixed-ratio schedule versus conditions of no reinforcement. This undifferentiated responding occurred even though a reinforcer assessment test was conducted just prior to very work session. All of these individuals responded with higher work rates on the fixed ratio schedule following the introduction of a device designed to make the contingency more discriminable. This research project will extend the previous research with these two groups. The individuals with profound retardation and other severe medical and sensory impairments who learned to respond differentially under contingent leisure-device activation will be exposed to a switch-feedback device configuration that permits to make choices across types of feedback. Individuals in this group who did not learn to respond differentially to the switch will be exposed to other switch-feedback device configurations in an attempt to produce exposed to experiments to develop communication responses using their device-supported responses. In summary, we propose to explore methods to (a) increase the proportion of persons with severe multiple disabilities who learn to control some aspect of their environment, (b) to enable individuals with rudimentary environmental- control repertoires to diversity their abilities to include forms of communication, and (c) to investigate how environmental supports shown to be effective in facilitating contingency learning can be used to establish communication.